On Thu, Sep 17, 2015 at 01:47:46PM +0000, Nick Nauwelaerts wrote:
> i would guess to join our nexus fex's in the pub, the also like to go missing
> in between rancid checkups.
>
> in our case it seems to be a wonky nx-os revision in combination with
> datacenter manager which seems to cause quite some load with its checkups.
>
> anything in the router's logs during the disappeances?
I've seen similar issues before with various hardware. Usually it's a
software bug
where two people are talking to the microcontroller at the same time
and there's no concurrency checking. We've exposed a lot of bugs by having two
scripts
do the same thing at the same time. Often a cisco device doesn't expect
concurrent memory/device access.
Recommendation:
Figure out how to make it happen, either in a tight loop, or having 3
windows
open doing while [1 == 1]: do clogin -x /tmp/rancid-commands hostname done
it might be as simple as finding the show controller or show inventory commands
and running those
in a loop. make sure cisco knows how you login and they reproduce it the same
way themselves with
these critical variables in mind:
1) via SSH
2) via IPv6
3) where SSH uses specific terminal types
4) where the SSH client offers keys
We had issues where optics would report odd things for a year or so and
filled a lot
of rancid logs. This was because Cisco wasn't expecting a certain older flavor
of their own
optic and their EEPROM validation code wasn't perfect.
- Jared
--
Jared Mauch | pgp key available via finger from [email protected]
clue++; | http://puck.nether.net/~jared/ My statements are only mine.
_______________________________________________
cisco-nsp mailing list [email protected]
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/
